Difference Amplifier/Common Mode Rejection Ratio

Brad Peirson

4-7-2005

EGR 214 – Circuit Analysis I

Laboratory Section 04

Prof. Blauch

Abstract

The purpose of this laboratory is to determine the Common Mode Rejection Ratio

(CMRR) for a difference amplifier circuit. A difference amplifier is a circuit that will

amplify the difference between two input voltages. A circuit schematic was given, and

from this schematic the relationship between the input voltages and the output voltages

was determined. The circuit was then drawn in P-Spice and simulated. This same circuit

was then built using a u741 operational amplifier chip. The CMRR was determined

algebraically, from the P-Spice model and from the results of measurements using a

digital multi-meter. The simulated and experimental results were then compared against

the CMRR from the mathematical model. While not ideal, the simulated and

experimental common mode rejection ratios were extremely close to the mathematically

predicted CMRR. Because of this the circuit will likely not show any variation in output

due to electrical interference.

1.0 Introduction

The difference amplifier is a circuit that requires two separate input voltages. The

purpose of the circuit is to amplify the difference between two input voltages. The circuit

also has the capability to effectively ignore the inputs when they are the same. This

function is vital as all electronic circuits are exposed to interference. Often such

interference will result in the input signals having the same voltage value. An ideal

difference amplifier will reject all such interference and amplify only the difference

between the two inputs.

The key to the difference amplifier is an operational amplifier. Since its inception

nearly sixty years ago the operational amplifier has been a key component in computer

systems. The internal circuitry in the modern operational amplifier is a complex circuit

consisting of resistors, capacitors and inductors. This complex circuit can be considered

a black box, which greatly simplifies the analysis of any operational amplifier circuit.

Section 2 outlines the operation of the operational amplifier and the difference

amplifier circuit. It also details several properties of the difference amplifier circuit.

Section 3 is the analysis of the difference amplifier circuit using assigned nominal

resistance values. This section also contains the derivations of the inherent relationships

- 1 -

of the difference amplifier. Section 4 contains a P-Spice simulation of the difference

amplifier circuit. Section 5 analyzes the data gathered from the physical circuit. Section

6 is the results discussion section and section 7 is the conclusions section.

2.0 The Operational Amplifier

The Operational Amplifier (OP AMP) is a complex integrated circuit. OP AMPS

are amplifying circuits that were originally designed to perform mathematical operations.

Early OP AMP consisted of vacuum tubes in early analog computers. As technology

evolved, so too did the OP AMP. The most common OP AMP configuration in use today

is the 741 integrated circuit chip. Figure 1 shows the 741 chip pin diagram with an OP

AMP schematic diagram.

vN +VCC

vP vO

-VCC

Figure 1: u741 OP AMP Pin Diagram

The OP AMP itself is a five pin device. It is also an active device that depends on

present inputs as well as previous inputs to operate. This means that the device requires a

voltage to drive it. This voltage is +VCC and –VCC, the same voltage with the polarities

reversed. These voltages are shown on the pin diagram in Figure 1. Pins 4 and 7 are

typically left off of schematic diagrams for the sake of clarity, though the voltages are

assumed to be present. The output voltage of the OP AMP cannot exist outside the ±VCC

range. When the output voltage would be less than –VCC, the output stays at –VCC. This

is known as the –Saturation region. Likewise, the output voltage remains at +VCC when

it would be greater. This is known as the +Saturation regions. Between the saturation

regions the OP AMP is in the linear operation region. The output voltage of the OP AMP

in the linear region is determined by

- 2 -

( )NPO vvAv −= . (1)

The relationship of the three regions is shown in Figure 2.

Figure 2: OP AMP Characteristics

The slope of the graph in the linear region is the voltage gain, A, from (1). The range of

vO can be used to determine the two main characteristics of an ideal OP AMP. The

relationship of vO to ±VCC is

CCOCC VvV +≤≤− . (2)

When (1) is substituted into (2) the relationship between vP, vN and ±VCC is found to be

( )AV

vvAV CC

NPCC +

≤−≤−

. (3)

The voltage gain is typically very high in an OP AMP, usually on the order of 105. In an

ideal OP AMP the gain is assumed to be infinite. When this is the case, the relationship

in (3) becomes

( ) 00 ≤−≤ NP vv . (4)

This shows the first characteristic of an ideal OP AMP. The second characteristic of the

ideal OP AMP has to do with currents. The resistance across the two input terminals on

an ideal OP AMP is assumed to be infinite. By Ohm’s law this means that the current

entering both input nodes must be zero. So, the characteristics of an ideal OP AMP are

- 3 -

NP vv = (5)

and

0== NP ii . (6)

2.1 Difference Amplifier

The difference amplifier is a specific application of the OP AMP. This circuit

amplifies the difference between two separate input voltages and ignores common inputs.

The schematic for this type of circuit is shown in Figure 3.

1R 2R

3R

4R

1v

2v

_

+

+ v_

0

Figure 3: Difference Amplifier Circuit

The relationship of the inputs to the outputs of the difference amplifier can best be

described if two new voltages are defined. The differential mode voltage and the

common mode voltage (v

)( dmv

cm) are defined as

12 vvvdm −= (7)

and

2

21 vvvcm+

= . (8)

These new voltages relate to the output voltages linearly. The new relationship is

shown in (9).

dmdmcmcmO vAvAv += (9)

- 4 -

When the two input voltages are the same, they drive the differential mode term to zero.

This means that the common mode voltage only shows up in the equation when the two

inputs are the same voltage with the same polarity. In order to ensure that this voltage is

rejected the common mode gain must be zero. When this happens it can be shown that

dmcmO vRRvv

1

20 += . (10)

This relationship shows that any common mode voltage will be disregarded. Likewise

any differential mode voltage will be amplified by1

2

RR

.

The relationship in (10) holds true only for an ideal difference amplifier. In

practical application there are no ideal circuits, however. The common mode gain and

the differential mode gain can be used to determine the common mode rejection ratio

(CMRR). The CMRR is a measure of how close the circuit is to being ideal and is stated

mathematically as

cm

dm

AA

CMRR = . (11)

A high CMRR is the overall goal because a large CMRR means that the common mode

voltage will have almost no impact on the output voltage. This is necessary because of

the nature of the common mode voltage. When the two input voltages have the same

value this is typically a result of electrical “noise.” A high CMRR means that this

interference will have very effect on the operation of the circuit.

3.0 Analysis

The overall equation from the schematic in Figure 3 can be used to verify the

relationship in (9). This is done using the superposition principle. First, is suppressed

to find the output with respect to v

2v

1. The resulting superposition circuit is shown in

Figure 4.

R1 R2

- 5 -

iN

vN v1 _

++ vO1 _

vP

R3

R4

Figure 4: First Superposition Circuit

Figure 1 showed that the voltage entering the negative node of the OP AMP was denoted

as vN and the voltage entering the positive node was vP. There is no vP in this

superposition circuit so by the relationship in (5) vN = vP = 0. KCL at the negative OP

AMP node can be used to determine the vO relationship for this circuit.

02

1

1

1 =−−

+−

NONN i

Rvv

Rvv

(12)

When the properties of the ideal OP AMP from (5) and (6) are applied to (12) the

relationship for the first output voltage can be found. This derivation is shown in (13),

(14) and (15).

02

1

1

1 =+Rv

Rv O (13)

1

1

2

1

Rv

RvO −= (14)

11

21 v

RRvO −= (15)

Once the relationship for vO1 is derived, the second superposition circuit can be

solved. This circuit is shown in Figure 5 below.

R1 R2

- 6 -

vN _

+vP +

vO2 _

R3

v2

R4

Figure 5: Second Superposition Circuit

By observing Figure 5 it can be seen that vN and vP can be found using voltage dividers.

These equations are

221

1ON v

RRRv+

= (16)

and

243

4 vRR

RvP += . (17)

Because OP AMPs have the property that vP = vN , (16) and (17) can be set equal to one

another and solved to find the relationship for the second output voltage. This full

derivation is shown in (18) through (21).

243

42

21

1 vRR

RvRR

RO +

=+

(18)

2431

2142 )(

)( vRRRRRRvO +

+= (19)

2

4

341

2

142

2

1

1v

RRRR

RRRR

vO

⎟⎠⎞⎜

⎝⎛ +

⎟⎠⎞⎜

⎝⎛ +

= (20)

2

4

31

2

12

2

1

1v

RRR

RRR

vO

⎟⎠⎞⎜

⎝⎛ +

⎟⎠⎞⎜

⎝⎛ +

= (21)

- 7 -

The final relationship for vO is the sum of the superposition solutions, and is

stated as

11

22

4

31

2

12

1

1v

RR

v

RRR

RRR

vO −⎟⎠⎞⎜

⎝⎛ +

⎟⎠⎞⎜

⎝⎛ +

= . (22)

At this point the relationship is in terms of the original input voltages. As shown

in section 2.1, the input voltages can be restated in terms of the common and differential

mode voltages. From (7) and (8) the values of the input voltages in terms of the mode

voltages were determined. These values are

dmcm vvv21

1 −= (23)

and

dmcm vvv21

2 += . (24)

These values can be substituted into (22) to determine the relationship between vP, vN and

vO. The complete derivation of this new relationship is shown below.

⎟⎠⎞

⎜⎝⎛ −−⎟

⎠⎞

⎜⎝⎛ +⎟⎟⎟⎟

⎠

⎞

⎜⎜⎜⎜

⎝

⎛

+

+= dmcmdmcmO vv

RRvv

RR

RR

RRv

21

21

1

1

1

2

4

3

2

1

1

2 (25)

1

2

4

311

12

1

2

4

311

12

22 RvR

RRRR

vRvRRvR

RRRR

vRvRv dmdmdmcmcmcm

O +⎟⎠⎞⎜

⎝⎛ +

++−

+

+= (26)

dmcmO vR

R

RRRR

RRv

RR

RRRR

RRv

⎟⎟⎟⎟

⎠

⎞

⎜⎜⎜⎜

⎝

⎛

+⎟⎠⎞⎜

⎝⎛ +

++

⎟⎟⎟⎟

⎠

⎞

⎜⎜⎜⎜

⎝

⎛

−+

+=

1

2

4

311

12

1

2

4

311

12

22 (27)

( ) dmcmO v

RRRRR

RRRRRR

v

RRRRR

RRRRRR

v⎟⎟⎟⎟

⎠

⎞

⎜⎜⎜⎜

⎝

⎛

+

++++

⎟⎟⎟⎟

⎠

⎞

⎜⎜⎜⎜

⎝

⎛

+

−−+=

4

3141

4

32212

4

3141

4

32212

2 (28)

dmcmO vRRRR

RRRRRRRRv

RRRRRRRR

v ⎟⎟⎠

⎞⎜⎜⎝

⎛+

++++⎟⎟

⎠

⎞⎜⎜⎝

⎛+−

=3141

32424142

3141

3241

22 (29)

- 8 -

( )( ) ( )

( ) dmcmO vRRR

RRRRRRv

RRRRRRR

v ⎟⎟⎠

⎞⎜⎜⎝

⎛+

++++⎟⎟

⎠

⎞⎜⎜⎝

⎛+−

=431

432214

431

3241

2 (30)

This final relationship fits the form shown in (9). The coefficient term on the

common mode voltage is Acm and the coefficient on the differential mode voltage is Adm.

The ideal difference amplifier will reject all common mode voltages. This only occurs

when R1=R3 and R2=R4. When this relationship is substituted into (30) the resulting

relationship matches (10).

Specific resistor values were assigned for use in this experiment. The nominal

resistor values used are shown in (31).

Ω=Ω=Ω=

Ω=

kRkR

kRkR

221231

4

3

2

1

(31)

Using these values for the resistors, R2 ≠ R4. This means that the difference amplifier in

this laboratory will not be ideal and there will be some common mode gain. A small

common mode gain is acceptable in the performance of the difference amplifier as long

as the difference mode gain is large. This will result in a large CMRR and will still allow

the circuit to operate with little interference from electrical “noise.”

The resistor values in (31) can be substituted into (30) to determine the common

mode and differential mode gains. These resulting gains are

045043.0−=cmA (33)

and

8789.22=dmA . (34)

The relationship of vO, vcm and vdm then becomes

dmcmO vvv 8789.22045043.0 +−= . (35)

The three-dimensional plot of this is shown in Figure 7.

- 9 -

Figure 7: 3-D Plot of vO using analysis data

Figure 7 shows that while (35) is not ideal, there is very little change due to vcm meaning

the impact it has on the equation is negligible. The CMRR for this model is

93464.507045043.08789.22

=−

=CMRR . (36)

Because this CMRR only involves nominal resistor values (36) will be the standard

against which the simulated and experimental results will be checked.

4.0 Simulation

The resistor values in (31) were input into P-Spice. There were two separate sets

of simulations run: one set where v1 = v2 and one set where v1 = -v2. For the first set the

vdm term in (9) became zero. For the second set the vcm term in (9) became zero. An

example of the P-Spice results for one of the simulations is shown in Figure 8.

- 10 -

Figure 8: P-Spice Simulation Schematic (v1 = v2 = 2V)

The results of the P-Spice simulations for v1 = v2 are given in Table 1. The common and

differential mode voltages were found by substituting the values of v1 and v2 into (7) and

(8). The common mode gain was calculated by substituting the values for vcm, vdm and vO

into (9).

- 11 -

Table 1: Common Mode Gain Simulation Results

v1 (V) v2 (V) vcm vdm vO (mV) Acm1 1 1 0 -47.73 -0.04773 2 2 2 0 -87.93 -0.043965 3 3 3 0 -132.12 -0.04404 4 4 4 0 -176.31 -0.044078 5 5 5 0 -220.49 -0.044098 6 6 6 0 -264.67 -0.044112

The data in Table 1 was input into a Microsoft Excel spreadsheet. The common mode

voltage was then plotted against vO and a linear regression was performed. This plot is

given in Figure 9.

y = -0.0436x - 0.0022R2 = 0.9998

-0.3

-0.25

-0.2

-0.15

-0.1

-0.05

00 1 2 3 4 5 6 7Vcm

Vo

Figure 9: vcm vs. vO for the simulated common mode gain

The common mode gain is the slope of the vcm vs. vO trend line. Figure 8 shows

that the common mode gain for the simulated results was -0.0436. This was the common

mode gain used in determining the relationship of vO to the common mode and

differential mode voltages.

After the simulations to determine the common mode gain, the simulations for the

differential mode gain were run. Again the values for vcm, vdm and vO were substituted

into (9) to determine the gain. The simulation results are shown in Table 2.

- 12 -

Table 2: Differential Mode Gain Simulation Results

v1 (V) v2 (V) vcm vdm vO (V) Adm-0.01 0.01 0 0.02 0.45997 22.9985 -0.02 0.02 0 0.04 0.91947 22.9868 -0.03 0.03 0 0.06 1.379 22.9833 -0.04 0.04 0 0.08 1.839 22.9875 -0.05 0.05 0 0.10 2.298 22.98 -0.06 0.06 0 0.12 2.758 22.9833

This data was again input into an Excel spreadsheet. The resulting plot with trend line is

given in Figure 10.

y = 22.98x + 0.0003R2 = 1

0

0.5

1

1.5

2

2.5

3

0 0.02 0.04 0.06 0.08 0.1 0.12 0.14

Vdm

Vo

Figure 10: vdm vs. vO for the simulated differential mode gain

Again, the slope of the plot shown in figure t is the differential mode gain. Both

of the gains determined in the regressions were input into (9) in order to determine the

relationship for the simulation data. This relationship was found to be

dmcmO vvv 98.220436.0 +−= . (37)

This equation was then graphed on a three dimensional coordinate system. The resulting

graph is shown in Figure 11.

- 13 -

Figure 11: 3-D Plot of vO using simulated data

The CMRR for the simulation data was

064.5270436.098.22

=−

=CMRR . (38)

5.0 Experimental Results

Four resistors with nominal values shown in (31) were chosen. The actual

resistances were measured and recorded in Table 3.

Table 3: Resistor Values

Resistor Nominal Value (kΩ) Measured Value (kΩ) R1 1 0.994 R2 23 22.764 R3 1 0.994 R4 22 21.74

The circuit in Figure 3 was built on the CADET circuit trainer using the resistors in Table

3. The first set of data measured was used to determine Acm. To do this the input

voltages were equal to one another. This makes the differential mode voltage zero. The

values were then substituted into (9) and solved for Acm. The results of these

measurements and calculations are given in Table 4.

- 14 -

Table 4: Common Mode Gain Experimental Results

v1 (V) v2 (V) vcm vdm vO (mV) Acm1.01 1.01 1.01 0 -0.038 -0.037624 2.00 2.00 2.00 0 -0.077 -0.0385 3.057 3.057 3.057 0 -0.101 -0.033039 4.08 4.08 4.08 0 -0.138 -0.033824 5.03 5.03 5.03 0 -0.179 -0.035586 6.08 6.08 6.08 0 -0.215 -0.035362

The experimental data was then input into an Excel spreadsheet. A regression

analysis was performed. The resulting plot is shown in Figure 12.

y = -0.0346x - 0.0021R2 = 0.9943

-0.25

-0.2

-0.15

-0.1

-0.05

00 1 2 3 4 5 6 7Vcm

Vo

Figure 12: vcm vs. vO for the experimental common mode gain

Again, the common mode gain will be the slope of the trend line in Figure 12.

The common mode gain for this experiment was -0.0346. The same type of process was

used for the differential mode data in Table 5.

- 15 -

Table 5: Differential Mode Gain Experimental Results

v1 (V) v2 (V) vcm vdm vO (V) Adm4.98 5 4.99 0.02 0.942 55.733 4.80 5 4.9 0.2 4.97 25.698 4.68 5 4.84 0.32 7.47 23.867 4.47 5 4.735 0.53 11.32 21.668 4.28 5 4.64 0.72 11.32 15.945 4.84 5 4.92 0.16 3.91 25.501

In this experiment, it was impractical to use equal but opposite voltages for v1 and

v2. There were two difficulties with the original set-up. First, if a single power supply

was used to provide the opposite voltages the supply would be shorted. This had the

potential of damaging the power supply. The second difficulty came if two supplies were

used to provide the opposite voltages. This meant that the power supplies would have

had to have been moved from one lab station to another. This method would have been

impractical. It was determined that since there was no practical way to provide the

necessary equal and opposite voltages, the voltages would be chosen sufficiently close to

one another so that the common mode voltage would be nearly constant. The gain could

be calculated using the common mode gain, the common mode voltage and the

differential mode voltage as in the previous data set. The data in Table 5 was again input

into Excel and a linear regression was performed. This regression plot is shown in Figure

13.

- 16 -

y = 15.574x + 1.5938R2 = 0.9289

0

2

4

6

8

10

12

14

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Vdm

Vo

Figure 13: vcm vs. vO for the experimental differential mode gain

The R2 value of the regression line in Figure 13 seemed too low for an accurate

best-fit line. The data was checked, and it was determined that the circuit was saturating

when the difference mode voltage was approximately 0.5. This was found to be the two

data points that resulted in vO=11.32. These two data points were left out of the

regression model, resulting in Figure 14.

- 17 -

y = 21.846x + 0.4999R2 = 0.9992

012345678

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35

Vdm

Vo

Figure 14: Modified vdm regression plot

The R2 value in Figure 14 was much higher than that in Figure 13 meaning that the

regression line is a better fit. The model in Figure 14 was the Adm value used for the vO

equation. The experimental vO equation was

dmcmO vvv 846.210346.0 +−= . (39)

This equation was then used in generating the three dimensional graph of vO. This graph

is given in Figure 15.

Figure 15: 3-D Plot of vO using experimental data

- 18 -

Finally, the CMRR for the experimental data was calculated in (40).

387.6310346.0846.21

=−

=CMRR (40)

6.0 Discussion

The CMRR based solely on the nominal resistance values in the circuit was

approximately 507. The CMRR calculated in section 4 was approximately 527, and the

CMRR from the experimental data was approximately 631. These values were all large

enough to conclude that outside electrical interference will not have any effect on the

operation of the circuit.

This is supported by the 3-D graphs in each section. Each of these plots showed

very little change due to the x-variable, vcm. This suggests that in each of the equations,

the common mode gain was nearly zero. In fact, when arbitrary values were chosen for

vcm and vdm it required an increase of nearly 1000 volts in vcm to create a noticeable

difference in vO.

The only discrepancy in the results of each model is in the CMRR for the

experimental data. This value was greater than the other two by approximately 100.

There were two possible explanations for this difference. The main reason was the

resistor values. The analysis and simulation CMRRs assumed that the resistances were all

equal to the nominal value. For the experimental CMRR, the resistances varied within the

allowable ±5% tolerance rating. This variance was likely large enough to cause a good

portion of the difference. The second reason for the difference between CMRRs was the

method in which the experimental vdm data was collected. In the ideal model, the

common mode gain is zero. In turn, since the common mode gain in the experimental

calculations was known to be non-zero the common mode voltage would have had to

have been zero. This proved impractical in the lab. Instead it was proposed that if v1 and

v2 were kept close enough together the common mode voltage could be assumed to be

constant. This assumption is the second source of error. The common mode voltage was

nearly constant but there was some amount of variation.

- 19 -

7.0 Conclusions

The difference amplifier circuit is a circuit that amplifies the difference between

two input signals. A mathematical model for the relationship of the output voltage to the

input voltage was derived in section 3.0. This model was used as the standard against

which the simulated and experimental results were checked. Because of the overall

precision in the three different common mode rejection ratios the results of the

experiment verify the characteristics of the ideal difference amplifier, also discussed in

section 3.0. In addition, the CMRR for all three sections was high enough to be

reasonably certain that electrical “noise” will not have any noticeable effect on the output

of the difference amplifier circuit.

- 20 -